SAYLES: THE SQUANTUM TILLITE. 165 
preserved as tillite, it must ordinarily be on a surface which is sub- 
siding at or soon after the time of the retreat of the ice-sheet. Any 
till deposit above sea-level on a stationary or rising surface would 
almost invariably be eroded long before later subsidence could remove 
it beyond the wear and tear of the elements. Whether the slate above 
the tillite is of marine or fresh water origin it is not possible at present 
to say. No clearly marine fossils have been found in it, and so far as 
this negative evidence goes it is more probable that this slate is of 
lacustrine origin. The absence of fossils, however, does not settle the 
question. Marine life in the Permian seas was scarce or wanting 
altogether in many places, and furthermore fossils are not found 
everywhere in the marine clays of Massachusetts and Maine and other 
places where marine clay of Pleistocene age outcrops. If volcanoes 
were situated then as now near the continental margins, the sea might 
not have been many miles away, for volcanic action’ was associated 
with the deposition of these beds as shown by melaphyre flows in 
several places in the Basin. According to Bailey Willis (1909, p. 
403-405) land extended at least 100 miles in a southeasterly direction 
from Boston and probably much farther than this. That there was 
high land to the southeast appears probable also from a study of the 
tillite. The evidence so far points to a southeasterly origin for the 
ice which formed the tillite. A discussion of this question of direction 
comes naturally in the history of the appearance of the tillite as shown 
best in the Atlantic exposure, and in a study of some features of the 
tillite found at the southeastern Squantum exposure. 
The Roxbury conglomerate proper at Atlantic exposes a thickness 
of about 520 feet. The lowest part shows rather small pebbles averag- 
ing about one inch in diameter. Farther up the pebbles increase in 
size gradually, while in the transition-beds below the tillite the pebbles 
are larger, averaging about four inches. It would seem very probable 
that this gradual increase in the size of the pebbles heralded the com- 
ing ice-sheet by wetter conditions or by a shorter distance from the 
source, as the ice drew nearer. If the larger size of the pebbles was 
due to more water and greater velocity, the pebbles should be as 
rounded as formerly, but if the approach of the ice was the cause of the 
size, the pebbles should be more angular as well as larger. The latter 
appears to be the case. 
Above the Roxbury a sandstone bed was formed, indicating slower 
stream-action. A bed of conglomerate was then laid down, indicating 
swifter stream-action. Another sandstone bed was then deposited. 
At this point a new phenomenon is met with. Above this last men- 
